Mixture of Auto-Associative Neural Networks for Speaker Verification
نویسندگان
چکیده
The paper introduces a mixture of auto-associative neural networks for speaker verification. A new objective function based on posterior probabilities of phoneme classes is used for training the mixture. This objective function allows each component of the mixture to model part of the acoustic space corresponding to a broad phonetic class. This paper also proposes how factor analysis can be applied in this setting. The proposed techniques show promising results on a subset of NIST08 speaker recognition evaluation (SRE) and yield about 10% relative improvement when combined with the state-of-the-art Gaussian Mixture Model i-vector system.
منابع مشابه
Factor analysis of mixture of auto-associative neural networks for speaker verification
This paper introduces the theory of factor analysis of the mixture of Auto-Associative Neural Networks (AANNs) with application in speaker verification. First, we formulate the problem of learning a low-dimensional subspace in part of the mixture of AANNs parameter space, and subsequently derive the update equations by minimizing loss function of the mixture. Second, we apply this technique to ...
متن کاملA Self-Reconstructing Algorithm for Single and Multiple-Sensor Fault Isolation Based on Auto-Associative Neural Networks
Recently different approaches have been developed in the field of sensor fault diagnostics based on Auto-Associative Neural Network (AANN). In this paper we present a novel algorithm called Self reconstructing Auto-Associative Neural Network (S-AANN) which is able to detect and isolate single faulty sensor via reconstruction. We have also extended the algorithm to be applicable in multiple faul...
متن کاملAdaptation transforms of auto-associative neural networks as features for speaker verification
We present a new approach of using Auto-Associative Neural Networks (AANNs) in the conventional GMM speaker verification framework with i-vector feature extraction and PLDA modeling. In this technique, an i-vector feature extractor is trained using adaptation parameters from a mixture of AANNs. In order to model parts of each speaker’s acoustic space, a training objective function based on post...
متن کاملAuto-Associative Neural Networks and Eigenbands Fusion for Frontal Face Verification
Face classification is an important area of research with many applications, including biometric security and searching face databases. This article describes an approach to verify faces using Auto-associative Neural Networks and Eigenbands fusion. In Eigenbands strategy each faces is divided in horizontal bands from which are extracted features using PCA. This method aims capture discriminativ...
متن کاملKernel-Based Probabilistic Neural Networks with Integrated Scoring Normalization for Speaker Verification
This paper investigates kernel-based probabilistic neural networks for speaker verification in clean and noisy environments. In particular, it compares the performance and characteristics of speaker verification systems that use probabilistic decision-based neural networks (PDBNNs), Gaussian mixture models (GMMs) and elliptical basis function networks (EBFNs) as speaker models. Experimental eva...
متن کامل